Monday, December 10, 2018

Saving Economics from Politics

An editorial appeared in The Hindu on November 30th which is reproduced here.

                                                                   Number Theory

"The larger lessons from the GDP back series must not be clouded by a political slugfest"


"Backcasting, or reworking past national accounts statistics based on the latest base year, is a regular exercise that governments carry out. Mainly done to enable precise comparison and analysis, it is a difficult exercise prone to contestation as it involves the inclusion of newer data sources, exclusion of outdated ones and making some subjective assumptions in the process. Throw in the political element, and GDP backcasting can become a controversial exercise, as it has now become in the case of the release of back series data from 2005-06 to 2011-12, the new base year. The data computed by the Central Statistics Office (CSO) and released by the Niti Aayog show that India never really grew in double-digits in 2010-11, nor was it the high-growth economy in the five years preceding this as earlier thought to be. It so happens that this period covers the two terms of the Congress-led United Progressive Alliance government, and the new data have predictably set off a political storm. The Congress may feel aggrieved as its biggest achievement, of taking India on the high GDP growth path, has come under question. During earlier instances of backcasting of GDP data, the political environment was not as deeply polarised as it is now, and so the exercise remained more academic.
The danger in the political slugfest now is that the many valuable insights that can be gleaned from the data will be lost sight of. The biggest of these is that India never really decoupled from the global economy during the years of the financial crisis (2008-10), unlike what was earlier believed. The new back series data show a much lower growth rate. This is an important learning for policymakers, going forward. Any criticism of the data has to take into account the fact that it has been generated by a thoroughly professional organisation, the CSO, and the methods have been scrutinised by experts, including past chief statisticians, and the Advisory Committee on National Accounts Statistics. Certainly, the release of the back series by the Niti Aayog goes against convention and is bad in optics. But this should not be reason to contest its integrity. The method of computation reflects the latest United Nations System of National Accounts; it also captures changes in the economy since 2004-05. Data sources have also been updated. Experts had testified to the robustness of the method when it was introduced in 2015, even while underlining that the availability of reliable data was crucial to arrive at the correct overall picture. There is little doubt that India needs to invest more in data collection and integration and do informal sector surveys more frequently. Robust, updated data are, in fact, insurance against politicians hijacking what is essentially an economic exercise."

    Comment on this editorial by the Readers' Editor of The Hindu on December 3rd is reproduced below:

"There were problems with the fundamental assumptions of a recent editorial on the new GDP back series

I generally refrain from commenting on editorials and opinion pieces. I recognise that there are points of convergence as much as there are points of divergence between the newspaper and its myriad readers, and even within the newspaper itself. These conversations lend plurality to the newspaper and they should not be viewed from any narrow ideological prism. However, I have to break from this norm to discuss the editorial “Number theory” (Nov. 30), which generated some sharp and divergent reactions.

Independence of the editorial

The arguments against the editorial were varied. Some took an ideological standpoint, while others interpreted the events that led to the release of the GDP back series. I would like to reiterate that my role as the Readers’ Editor is not that of a pre-publication censor, but of a post-publication evaluator. I do get complaints about The Hindu’s editorial policy, which is defined by the editor and his editorial team. I can explain the policy but I cannot interfere with it. It is vital to support the independence of the editorial. The acid test for the Readers’ Editor is how he conducts himself when his own opinion is at variance with that of the paper. Can he be an effective advocate for free speech, tolerance and plurality if he lacks these democratic traits? Hence, the issue I am discussing is not about the ideological thrust of the editorial but its fundamental assumptions.
The assertion of the editorial that “robust, updated data are, in fact, insurance against politicians hijacking what is essentially an economic exercise” seems like a statement yearning for an ideal reality rather than one that is based on reality. The sequence of events since the creation of Niti Aayog seems to point at a complete politicisation of numbers. Some facts lend credence to the criticism of the numbers put out by Niti Aayog. One, the government had failed to appoint a Chief Statistician for nearly eight months after the retirement of T.C.A. Anant in January. Two, it has had a tense relationship with the Reserve Bank of India. Three, a set of data presented by the committee set up by the National Statistical Commission was withdrawn. Four, the Agriculture Ministry backtracked on a report showing the adverse effects of demonetisation on farmers. Five, there’s the timing of the new data, which many see as a desperate ploy to distract people’s attention from the trenchant criticism of demonetisation by the former Chief Economic Adviser. The line between the Central Statistics Office and Niti Aayog is blurred, thereby lending a political colour to an essential economic exercise.

The practice of data torture

If the editorial is read along with the explainer “What’s with the back series GDP data?” (December 1), it is clear that the editorial jumped the gun to grant the benefit of doubt to the latest exercise and suspended essential journalistic curiosity. The explainer deals with the problem of finding matching data for the older series to the present MCA-21, which is available only since 2011-2012. As a journalist, my entry points for understanding a range of subjects have been science and literature. About pure qualitative methods and number crunching, one of the finest historians of science, Thomas Kuhn, observed that “nature undoubtedly responds to the theoretical predispositions with which she is approached by the measuring scientist.” The Anglo-American economist, Ronald Coase, gave an interesting economic reading of this statement: “If you torture the data long enough, it will confess to anything.” It is true that governments and institutions do indulge in data torture, a practice of repeatedly interpreting source data until it reveals a desired result.
The editorial seems to be oblivious to this caution from the Nobel laureate. Its statement that “during earlier instances of backcasting of GDP data, the political environment was not as deeply polarised as it is now, and so the exercise remained more academic” fails to capture the full picture. In all the earlier changes, the methodology was not only robust but also transparent, and incomparable parameters were hardly used to deduct a number to understand macroeconomic reality. This was an editorial of forking paths, for how do we reconcile the assertion that “the release of the back series by the Niti Aayog goes against convention and is bad in optics” with the conclusion that “this should not be reason to contest the integrity of the new numbers”?
readerseditor@thehindu.co.in"

           My response to the comment of the Readers' Editor
Dear Sir,

I refer to your column in The Hindu today.

Though you are hesitant to comment on editorials, I think you have every right and perhaps even duty to make objective observations on anything that concerns readers. Readers do read, nay even study, the editorials in a newspaper like ours.

As a former banker, I have very often observed that a manager averse to lending ultimately selects a wrong borrower to lend.  In a similar way, though you are averse to comment on an editorial, you have finally chosen a wrong editorial to criticise.

You have noted the assertion that "robust, updated data are , in fact, insurance against politicians hijacking what is essentially an economic exercise" is a statement yearning for an ideal reality (sic) rather than one that is based on reality. In the process, you have ignored the presence among readers of those who look forward to The Hindu to play an effective role in shaping the tone and tenor of economic discussions. Freeing economic discussions from political prejudices is not only desirable but also essential to promote economic development. That The Hindu has done this in the editorial under reference despite the newspaper's proclivity to be harsh on the present powers-that-be is welcome. Please don't dissuade the editors from objective analyses.

The editorial has rightly drawn distinctions between who prepared the data and who communicated them to the wider populace. We should not stigmatise the data because the communicator (Niti Aayog) is allegedly political. The leader has correctly referred to the "fact that it has been generated by a thoroughly professional organisation , the CSO, and the methods have been scrutinised by experts ------". What else do you desire?

I salute The Hindu for its brave attempt to elevate economic discussion from the cesspool of dirty and partisan politics. I hope that the newspaper will continue this crusade despite the dominant presence of carping critics with a hidden political agenda.

Regards,
K.R.Srivarahan,

No comments: